FULL STORY

OpenAI's Claimed Refutation of the Exponential Time Hypothesis: From Prank Rumor to Paper

What began as a suspected prank rumor about models refuting the Exponential Time Hypothesis escalated after OpenAI published a paper claiming exactly that, spurring derived subexponential algorithms.

2026-10-27 ~ 2026-10-29 · 3 episodes · 9 posts

Episode 1 · Viral Claims That Claude and GPT-Next Refuted the Exponential Time Hypothesis Appear to Be a Hoax (2026-10-27, 4 posts)

Posts circulating on Reddit and elsewhere claim Claude and GPT-Next simultaneously produced 2^o(n)-time algorithms for 3-SAT, refuting the Exponential Time Hypothesis, but the claims appear to stem from a spoofed paper cover. A separate Hexagon-hosted, "mostly AI-generated" paper claims an internal Anthropic model autonomously discovered such an algorithm using 14 million output tokens without human input.

Episode 2 · Subexponential Hamiltonian Cycle Algorithm Built on ETH Refutation (2026-10-28, 3 posts)

Building on OpenAI and Anthropic's recent refutation of the Exponential Time Hypothesis, a researcher derived and Lean-verified a deterministic subexponential algorithm for exactly counting Hamiltonian cycles in O(n)-edge graphs.

Episode 3 · OpenAI Paper Claims the Exponential Time Hypothesis Is False (2026-10-28, 2 posts)

OpenAI's preprint titled "The Exponential Time Hypothesis Is False" claims a subexponential deterministic algorithm for 3-SAT, drawing viral attention and skeptical amusement from the community.